Sampled-data strategies for the control of a 3 DOF hover system

Abstract

This thesis will investigate different sampled-data strategies that can be used for the digital implementation of a controller. To implement the controller several discretization techniques are considered and compared including implicit and explicit Euler methods, Tustin approximation and zero-order hold. The first sampled-data strategy to be investigated is classical periodic sampling, where the output signal is measured periodically, and the controller output is updated every fixed amount of time. The deterioration of performance as sampling times increase as well as the stability properties is investigated for the different discretized models. In a more advanced stage of this project, an event-triggered strategy is considered. In event-triggered control, sensor sampling and control updates are generated only when the system’s state deviate too much from the desired value, which offers a reactive approach to control.

This thesis will investigate different sampled-data strategies that can be used for the digital implementation of a controller. To implement the controller several discretization techniques are considered and compared including implicit and explicit Euler methods, Tustin approximation and zero-order hold. The first sampled-data strategy to be investigated is classical periodic sampling, where the output signal is measured periodically, and the controller output is updated every fixed amount of time. The deterioration of performance as sampling times increase as well as the stability properties is investigated for the different discretized models. In a more advanced stage of this project, an event-triggered strategy is considered. In event-triggered control, sensor sampling and control updates are generated only when the system’s state deviate too much from the desired value, which offers a reactive approach to control.